Linguistic relativity (Sapir-Whorf hypothesis)

Overview

Linguistic relativity is a theoretical relationship between human language and thought, holding that a speaker's native language influences his or her cognition. The strong form of this theory holds that language is deterministic; in the weak form, language merely has a powerful influence on thought. The theory is often referred to as the Sapir-Whorf hypothesis, especially in introductory linguistics texts or news media, but this is a misleading term for several reasons: Neither Sapir nor Whorf originated the idea, nor did they work together; the most common forms of the idea were developed by later theorists; and neither man framed the idea as a hypothesis, which by definition must be testable. Because Whorf popularized the idea of linguistic relativity, schools of thought arguing in favor of it (especially in its weak form) are called Whorfian.

Ideas reflecting the principle of linguistic relativity are some of the most well-known ideas about linguistics among laypeople. News stories regularly circulate recapitulating an idea first put forward by anthropologist Franz Boas: that Inuit and Yupik languages have an unusually high number of words for snow, reflecting the significant role snow plays in their day to day lives. As discussed in greater detail below, while this is half-true, the relativistic explanation is questionable. Nevertheless, a great many variations on "culture X has many words for Y" or "culture P has no words for Q" are published in the popular press, in many cases misrepresenting or exaggerating the findings of linguistics research.

Whorfianism has a particular popularity outside of linguistics, whether because it is a compelling "just-so story" about language or because it continues to be taught in many introductory linguistics textbooks. For example, in 2012, Yale economist Keith Chen invoked a strong Whorfian view in a working paper on grammar and risk aversion. Chen argued that speakers of languages with a clear grammatical future tense marking evinced little concern for the future, and consequently had higher rates of obesity, smoking, and debt, and showed poor financial planning for the future. On the other hand, speakers of languages that use present-tense forms for future time references were more likely to have healthy eating habits and good financial planning. As many linguists pointed out, the most immediate flaw in Chen's reasoning was the Piraha of Brazil, whose language has no future tense marking and whose culture is notable for its extreme disregard of planning for the future, to the extent that they are one of the few hunter-gatherer cultures with no food preserving traditions.

Further Insights

The first proposals of linguistic relativity were offered in the nineteenth century. Modern linguistics developed in the eighteenth and nineteenth centuries, primarily working on Indo-European languages and especially the reconstruction of Proto-Indo-European, but there was considerable overlap in the nineteenth and early twentieth centuries between linguists and anthropologists. This interest in other cultures and their languages naturally gave rise to the notion that linguistic differences and cultural differences could be connected. The early nineteenth century Prussian linguist, philosopher, and diplomat Wilhelm von Humboldt (brother of the famous naturalist Alexander von Humboldt) proposed that linguistic differences represented a nation's Weltanschauung (worldview) and more specific formulations along the same lines were offered by later linguists and anthropologists. Anthropology at the time was often deeply inflected by racist, colonialist, and imperialist ideas; as a result, the question addressed was sometimes not "what do these differences in language say about these groups of speakers," but "why is this group of speakers inferior to this other group (which is to say, inferior to us)?" Boas stood out in stressing the equality of all cultures and languages, making it all the more ironic that his work would be put to use to "explain" the inferiority of non-European cultures.

The Sapir-Whorf hypothesis is named for Boas's student Edward Sapir and Sapir's student Benjamin Lee Whorf. Sapir believed that differences in the grammatical systems of languages ultimately indicated that native speakers of two different languages perceived reality in different ways—the first full-throated expression of linguistic relativity. However, Sapir also repeatedly argued against linguistic determinism, and much of his work demonstrates that while he believed those differences in perception must exist, he did not believe the relationship between language and culture was especially strong. It is Sapir's student Whorf who became the real father of what he called "the linguistic relativity principle." Though he studied with Sapir, opponents of Whorfianism point out that he was an amateur enthusiast of linguistics, who never earned an advanced degree; he was an insurance man, employed by the Hartford Fire Insurance Company. Whorf drew on Boas's "words for snow" example, exaggerating it beyond any claim Boas had made, in an attempt to argue in favor of linguistic relativity. It was one of several examples of polysemy in indigenous languages that he used to show that those languages made more granular distinctions than European languages. He also argued that Hopi speakers experienced time in a fundamentally different way than English speakers, which was reflected by—or caused by—their language. Later scholars have questioned Whorf's evidence.

More than anyone else, Whorf introduced linguistic relativity into linguistics discourse, despite never formulating a hypothesis. Later linguists proposed specific hypotheses, though, which Roger William Brown summarized as the weak form (structural differences between languages are paralleled by nonlinguistic differences in cognition of the native speakers) and the strong form (structural elements of a speaker's native language partially or fully determine his or her worldview). The strong form, or linguistic determinism, never attracted much support, and professional linguists have almost unanimously rejected it. The weak form has varied in its popularity. It remains commonly taught in introductory courses, and so many people who are not linguists themselves but took a linguistics or linguistics-related class in college have been exposed to it, which may be its main source of popularity. Universalists, beginning with Noam Chomsky's proposal of Universal Grammar, argue that all languages share the same underlying structure—that the structural differences Whorfians are talking about are superficial compared with the deep structures. Universalism dominated the academy from the 1960s to the 1980s, and was popularized by Stephen Pinker's popular science book The Language Instinct (1994).

Defenders of linguistic relativity tend to come from cognitive linguistics, an interdisciplinary study of linguistics and psychology. George Lakoff is among the best known of these new Whorfians, and is known for his conceptual metaphor thesis, which emphasizes the mind's use of metaphors to apprehend phenomena. In Lakoff's formulation, different languages employ different cultural metaphors, and the differences in those metaphors shed light on differences in cognition. Arguably Lakoff's defense is far afield from the arguments made by the mid-century Whorfians, though he argues that Whorf was discredited largely by misrepresenting the Whorfian position.

Since the 1980s, empirical research in linguistic relativity has given it new credibility by generating quantifiable results supporting the weak form. Domain-centered research, for example, compares semantic domains across linguistic groups, such as color terms. While color perception is determined by the neural system, how speakers distinguish shades seems to be driven by language. A common example is the blue-green distinction, especially since early research on color terms indicates that only languages with a blue-green distinction will develop terms for purple, pink, orange, grey, or brown (Cibelli, Xu, Austerweil & Regier, 2016).

Issues

A key concept in the discussion of linguistic relativity is that of untranslatability: the property of a text or utterance such that no translation in another language can be found. Untranslatability is a tricky concept. There are numerous lists of supposedly untranslatable words that circulate on the Internet and are occasionally picked up by the popular press. The problem with this is that the list is accompanied by definitions of the words … which clearly constitute translations. Whether the word is saudade (a Portuguese word for nostalgia for something lost) or flaneur (a French term for someone who takes a leisurely stroll without a destination), at best these words are words that cannot be translated into a single English word. This is far from remarkable: Some of the most basic utterances in Romance languages require more than one word when translated into English, like digame in Spanish ("tell me"). Furthermore, the existence of loan words calls untranslatability into question at the word level: there is no native English word for "taco," but does this tell us anything meaningful, or impact English speakers' ability to speak and think about tacos?

For something to really be untranslatable, it must be true that in the best approximation of that utterance or text when translated into another language, something non-trivial is lost. This is a perennial issue in professional translation. The direct equivalent of a word may have a connotation that the original does not, for example. This comes up frequently in literary translations in a number of recurring areas. Swear words differ so much from language to language that the direct translation of a swear word in one language may not be a swear in another, or the translation of a mild swear word in one language may be considered much more obscene in another (Bowers & Pleydell-Pearce, 2011). Puns and other wordplay prove difficult to translate, not because the words are difficult but because the text as a whole depends on similarities, rhymes, or other relationships among words that may not be retained through translation.

Further, not all languages make the same distinctions in gender: English, for example, uses separate words for "aunt" and "uncle," but only one word for "cousin," regardless of gender. When translating the English "cousin" into a language that has separate words for male or female cousins—or even separate words for first, second, or third cousins, or for cousins and cousins-once-removed—the translator may need more information than is available in the text. Complementing that dilemma, English has no term for the relationship between the parents of one spouse and the parents of the other spouse in a marriage. The mother of your wife is your mother-in-law; your father is your wife's father-in-law. But what is your mother-in-law to your father? English is mum, but many other languages have terms for these relationships. While the concept can be conveyed in English, it is cumbersome to translate them. The question, though, is whether this means that non-English speakers think about their relatives differently than English speakers do, and whether their language is responsible for this difference.

Franz Boas is especially known for his assertion that the Eskimo-Aleut family of languages has an unusually high number of words for snow. Over the decades—Boas first proposed the idea in the 1910s—this idea trickled into the popular consciousness as one of the best-known ideas about language, generally received as proof that environment or culture dictate language in ways that in turn influence thought: that is, Inuit people deal with snow more often than, say, Romans, and so have an unusually high number of words in order to make finer distinctions between types of snow or snow phenomena than other languages do. Perhaps the main reason the idea became so well-known is that Whorf used it, in 1940, to support his ideas about linguistic relativity; Boas himself had only been discussing different ways languages form words, though that point has commonly been misunderstood.

There are a number of problems with these ideas. First, it is not clear that Boas's assertion (or at least Whorf's representation of it) is true in any meaningful sense: the Dictionary of West Greenlandic Eskimo Language lists only two root words for snow, one for snow in the air and the other for snow on the ground. But even beyond that—and even before dealing with the relativistic or deterministic claim the idea is used to promote—the idea that "X has many words for Y" is one that seems straightforward to laypeople, but is much less so to linguists. In English, is "snow flurry" a separate word from "snow"? There is a sense in which "snow flurry" is clearly not "a word" in the way we usually mean it—it is a two word phrase, not a single word—but this is not necessarily because English has few words for snow, but rather because multi-word phrases are one of the ways English constructs vocabulary. While "snowstorm" and "thundersnow" are written as compound words, it is a purely arbitrary distinction that puts them in a separate category as "snow flurry." This is a minor example of the problem with proving "words for snow": It is not easy to define what we count as a "word," nor how to count words or phrases derived from other words.

The relativistic claim depends not so much on how many words a language has for snow (or anything else) but how many types of snow and snow phenomena that language can distinguish between. If you were to assemble a list of types of snow and snow phenomena—wet snow, powdery snow, falling snow, snowbanks, slush, snowflake, and so on—and two languages have terms for the same number of items on the list, does it matter how those languages construct those terms? English has one of the largest vocabularies in the world, but it would not make sense to suggest that this reflects the greater familiarity of native English speakers with all the many areas of life touched on by that vocabulary. Instead it reflects a difference in the ways languages function and construct their vocabularies.

Linguists like Geoffrey Pullum have pointed out that the "words for snow" narrative plays in to a desire to "other" or exoticize other cultures. In his piece "The Great Eskimo Vocabulary Hoax" (1991), he argues the narrative "comports so well with the many other facets of their polysynthetic perversity," like diets and social customs unfamiliar to Westerners. Pullum also tracks how the number of words for snow attributed to the Inuit has grown over the decades as the "words for snow" story is told.

The "words for snow" story has become a common example used to allude to linguistic relativity: a 2003 article in The Economist invoked it stating, "If Eskimos have dozens of words for snow, Germans have as many for bureaucracy." Its commonality has even resulted in inspiring the word "snowclone" for a particular kind of clichéd phrasal template. Other snowclones include "the mother of all X," "X is the new Y," and "have X, will travel."

Bibliography

Athanasopoulos, P., Bylund, E., & Casasanto, D. (2016). Introduction to the special issue: New and interdisciplinary approaches to linguistic relativity. Language Learning, 66(3), 482–486. Retrieved March 15, 2018, from EBSCO Academic Search Ultimate. http://search.ebscohost.com/login.aspx?direct=true&db=asn&AN=117017759&site=ehost-live

Athanasopoulos, P., Casaponsa, A. (2020, May 31). The Whorfian brain: Neuroscientific approaches to linguistic relativity. Cognitive Neuropsychology, 37, 393-412, https://doi.org/10.1080/02643294.2020.1769050

Bender, A., Beller, S., & Klauer, K. C. (2016). Lady Liberty and Godfather Death as candidates for linguistic relativity? Scrutinizing the gender congruency effect on personified allegories with explicit and implicit measures. Quarterly Journal of Experimental Psychology, 69(1), 48–64. Retrieved March 15, 2018, from EBSCO Academic Search Ultimate http://search.ebscohost.com/login.aspx?direct=true&db=asn&AN=112333070&site=ehost-live

Bowers, J. S., & Pleydell-Pearce, C. W. (2011). Swearing, euphemisms, and linguistic relativity. Plos ONE, 6(7), 1–8. Retrieved March 15, 2018, from EBSCO Academic Search Ultimate. http://search.ebscohost.com/login.aspx?direct=true&db=asn&AN=74548859&site=ehost-live

Casasanto, D. (2016). A shared mechanism of linguistic, cultural, and bodily relativity. Language Learning, 66(3), 714–730. Retrieved March 15, 2018, from EBSCO Academic Search Ultimate. http://search.ebscohost.com/login.aspx?direct=true&db=asn&AN=117017761&site=ehost-live

Cibelli, E., Xu, Y., Austerweil, J. L., Griffiths, T. L., & Regier, T. (2016). The Sapir-Whorf hypothesis and probabilistic inference: Evidence from the domain of color. Plos ONE, 11(7), 1–28. doi:10.1371/journal.pone.0158725

Feist, M. I. (2016). Minding your manners: Linguistic relativity in motion. Linguagem Em (Dis)Curso, 16(3), 591–602. Retrieved March 15, 2018, from EBSCO Academic Search Ultimate. http://search.ebscohost.com/login.aspx?direct=true&db=asn&AN=120387700&site=ehost-live

Frothingham, M. B. (2023, September 1). Sapir–Whorf Hypothesis (Linguistic Relativity Hypothesis). Simply Psychology. www.simplypsychology.org/sapir-whorf-hypothesis.html

Lucy, J. A. (2016). Recent advances in the study of linguistic relativity in historical context: A critical assessment. Language Learning, 66(3), 487–515. Retrieved March 15, 2018, from EBSCO Academic Search Ultimate. http://search.ebscohost.com/login.aspx?direct=true&db=asn&AN=117017760&site=ehost-live

Scholz, B. (2024). Whorfianism. Stanford Encyclopedia of Philosophy. plato.stanford.edu/entries/linguistics/whorfianism.html

Sharifian, F. (2017). Cultural linguistics and linguistic relativity. Language Sciences, 59, 83–92. Retrieved March 15, 2018, from EBSCO Academic Search Ultimate. http://search.ebscohost.com/login.aspx?direct=true&db=asn&AN=120232476&site=ehost-live

Sidnell, J., & Enfield, N. J. (2012). Language diversity and social action: A third locus of linguistic relativity. Current Anthropology, 53(3), 302–333. Retrieved March 15, 2018, from EBSCO Sociology Source Ultimate. http://search.ebscohost.com/login.aspx?direct=true&db=sxi&AN=75172903&site=ehost-live